Transcriptome analysis identified the mechanism of synergy between sethoxydim herbicide and a mycoherbicide on green foxtail

Green foxtail (Setaria viridis, GFT) is a globally important weed and abundant on the Canadian prairies. Sethoxydim, an ACCase-inhibiting herbicide used to manage GFT, faces reduced efficacy due to resistant biotypes. Previous work showed a host-specific fungal pathogen, Pyricularia setariae, is efficacious against GFT and that sublethal sethoxydim application can synergize fungal biocontrol, extending efficacy to older plants and other foxtails. However, the physiological and molecular mechanisms underlying this synergy were unknown. Understanding these chemical–microbial interactions can inform novel weed biocontrol strategies. Leveraging genomic resources (with foxtail millet, Setaria italica, as a related reference), this study constructed de novo reference transcriptomes for herbicide-sensitive (HS) and herbicide-resistant (HR) GFT biotypes and used RNA-seq to explore the mechanism of synergy between sublethal sethoxydim and P. setariae.

Synergy between synthetic herbicides and mycoherbicides has been observed across systems. Co-application of Colletotrichum coccodes with thidiazuron enhanced control of velvetleaf compared to either alone; sublethal 2,4-D+MCPP improved biocontrol of field bindweed by Phoma proboscis; and glyphosate increased infection success of Alternaria cassiae in sicklepod. It has been suspected that certain herbicides suppress plant defense responses, thereby facilitating fungal infection and improving weed control. Despite such reports, the molecular mechanisms of these synergistic interactions, particularly in weedy species, remain poorly characterized. The lack of weed transcriptomic resources has likely limited mechanistic insights, motivating the de novo assembly and transcriptome-guided analysis performed here.

Plant and pathogen materials: Two GFT biotypes were used: an ACCase-inhibitor-sensitive (HS) and a resistant (HR) biotype collected from Manitoba, Canada, species-verified, and with HR confirmed by sethoxydim application at 1x–3x label rate. Seeds were increased in greenhouse. Growth conditions included Sunshine #3 potting mix with 1% controlled-release fertilizer in 15-cm pots, 17–20 °C greenhouse, 14-h supplementary light (~230 µE m−2 s−1). Seven-day-old seedlings were also produced on wet filter paper. For exogenous ABA testing, HS seedlings germinated on paper were transplanted to pots at 25 plants per pot. Pathogen and inoculum: Pyricularia setariae isolate 94-409A (highly virulent to both biotypes) was cultured on oatmeal agar at 26 °C under near-UV, spores harvested into water, adjusted to 1×10^8 spores/ml, stored at 4 °C, and amended with 0.1% Tween 80 prior to inoculation. Synergy experiment: Four treatments per biotype in a randomized complete block with four replicates (pots): mock (water), herbicide only, P. setariae only, and herbicide + P. setariae. At the 3-leaf stage, plants received 0.1× sethoxydim (Poast Ultra: 0.3 ml/L water + 0.01% Merge adjuvant), applied at 10 ml/m^2 (~50 cm distance) to produce slight damage in HS. Twenty-four hours later, plants were inoculated with ~3 ml spore suspension per pot to near runoff using an airbrush at 275 kPa. After 24 h in a dew chamber (20±2 °C, dark), plants were returned to greenhouse. Leaves were sampled at 2 dpi (three top leaves) for RNA extraction. Weed control was assessed 4–7 dpi; fresh biomass per pot was measured at 7 dpi by cutting plants at soil level and weighing combined healthy and dead tissues. Reference transcriptome development: For each biotype, top leaves were collected at 7, 24 (3-leaf), 42, and 67 days after emergence. Five plants per pot were bulked per biological replicate; three replicates per stage. Total RNA was extracted (RNeasy Plant Mini Kit with on-column DNase), quality-checked (RQI>9). Equimolar RNA from stages was pooled per biotype to construct RNA-seq libraries (TruSeq RNA v2). Sequencing was on Illumina MiSeq (V3, 150 cycles). Reads were quality-filtered (remove reads with >5% Ns or >50% bases with Q<5). De novo transcriptome assemblies (CLC Genomics Workbench v8.5) were optimized across 30 parameter combinations; final parameters were k-mer 64, length fraction 0.95, similarity fraction 0.95 to minimize chimeras and maximize BLAST hits. Assembled contigs were queried via BLASTx (e≤0.01) against Setaria italica proteins (Phytozome v10.1) and also compared to Panicum virgatum, Zea mays, and Sorghum bicolor protein datasets (e≤0.05, similarity ≥70%). Contigs without significant hits were compared to NCBI nr. RNA-seq for treatments and DEG analysis: For each biotype and treatment (control, herbicide, fungus, herbicide+fungus), RNA-seq reads (post-filter) were mapped to the respective de novo transcriptome. Expression was quantified as RPKM. Differential expression was determined via Empirical Analysis of DGE with thresholds |log2FC|≥2 and FDR≤0.01. PCA was used to examine replicate clustering. Functional annotation of contigs used Blast2GO Pro (BlastX against NCBI nr with Viridiplantae filter). GO enrichment (Fisher’s Exact Test via Gossip) contrasted UP-DEGs (test) versus DOWN-DEGs (reference), with FDR≤0.01, retaining the most specific significant GO terms. RT-qPCR validation: Twenty transcripts were selected from DEGs (photosynthesis genes and ABA-activated signaling components including bZIP60) for validation. cDNA was synthesized from 1 µg RNA (SuperScript III). qPCR employed SYBR Green chemistry (StepOne Plus) with melt-curve and gel checks; actin served as the endogenous control. Three biological and three technical replicates were used. Log2 fold-changes from qPCR were compared to RNA-seq. ABA effect assay: To test ABA’s role, HS plants received a soil drench of 500 µM ABA at the 3-leaf stage, 24 h before inoculation. ABA stock (50 mM) was prepared in 1 N NaOH and diluted; a 0.01 N NaOH drench served as vehicle control. Treatments mirrored the synergy experiment. Leaves at 2 dpi were collected for RT-qPCR of bZIP60 (Contig_GFT-HS_9248). Statistics: Experiments were repeated and pooled after Shapiro–Wilk normality and Bartlett’s homogeneity tests. ANOVA (SAS v9.3) with LSD at P<0.05 separated treatment means. For qPCR, ANOVA and Fisher’s LSD (P<0.05) compared expression levels.

- Synergy in HS but not HR: At 0.1× label rate, sethoxydim caused slight injury in HS but not HR plants. P. setariae infection alone caused necrosis in both, being somewhat more aggressive on HR. The herbicide + fungus combination substantially increased injury and reduced fresh weight only in HS; there was no added effect in HR. - Transcriptome assemblies: HS and HR de novo assemblies yielded 29,515 and 29,137 contigs, respectively, with total assembled lengths of 23.65 Mb (HS) and 22.21 Mb (HR); N50 values were 1263 bp (HS) and 1182 bp (HR). Approximately 89.1% (HS) and 88.8% (HR) of reads mapped back to assemblies. About 85% of contigs had BLAST hits in NCBI nr; most top hits were to Setaria italica. Comparative analyses indicated high homology to S. italica and fewer to P. virgatum, S. bicolor, and Z. mays. - RNA-seq for treatments: Mapping covered ~93% of assembled contigs across treatments. DEG counts (|log2FC|≥2, FDR≤0.01): HS—herbicide 1,667; P. setariae 7,654; herbicide+P. setariae 8,141. HR—herbicide 46; P. setariae 8,002; herbicide+P. setariae 6,970. PCA showed herbicide altered HS transcriptome but not HR. - GO enrichment: In HS, 139 GO terms were enriched (63 up, 76 down); in HR, 36 up and 76 down. Photosynthesis-related processes (e.g., thylakoid membrane organization, chlorophyll biosynthesis, photosynthetic electron transport) were strongly down-regulated in HS under herbicide and pathogen treatments; HR showed downregulation mainly under pathogen. - ABA signaling and ubiquitination: ABA-activated signaling and protein ubiquitination were uniquely up-regulated in HS under herbicide+pathogen synergy. NCED (ABA biosynthesis) gene (HS contig 16797) showed ~150-fold upregulation. bZIP transcription factors increased, notably bZIP60 (HS contig 9248) with ~7.1-fold change; expression was higher in herbicide and herbicide+pathogen than pathogen alone. Ubiquitination components (E1 activating enzymes: HS contigs 16970, 3695; E2 conjugating enzyme: contig 4091; E3 ligases: contigs 1140, 25781, 4878; U-box proteins) were up-regulated ≥2-fold in HS synergy. - Photosynthesis gene behavior: C4 photosynthesis genes in HS were down-regulated by herbicide, pathogen, or synergy; in HR, the herbicide tended to up-regulate these genes while pathogen/synergy down-regulated them, suggesting HR tolerance involves insensitivity or compensatory activation of photosynthetic genes. - ABA phenocopies synergy: Exogenous ABA (500 µM drench) modestly reduced growth and markedly enhanced P. setariae infection and biomass reduction, comparable to herbicide+fungus in HS. ABA elevated bZIP60 expression to levels similar to herbicide. Combining ABA with herbicide did not further increase bZIP60 expression, indicating action through a shared pathway. - Validation: RT-qPCR fold changes correlated with RNA-seq for selected genes, supporting data reliability.

The study addressed how sublethal sethoxydim synergizes P. setariae-mediated biocontrol of GFT. Molecular profiling revealed that in HS plants, herbicide predisposition combined with pathogen challenge activated ABA-dependent signaling and bZIP60, concomitant with suppression of photosynthesis-related processes. The IRE1/bZIP60 branch of the unfolded protein response (UPR) is implicated in modulating plant defense; elevated bZIP60 likely attenuates defense responses, facilitating pathogen colonization and enhanced biocontrol. Protein ubiquitination machinery was also up-regulated, suggesting increased protein turnover associated with stress and defense modulation. In HR plants, the herbicide had negligible transcriptomic impact, explaining the absence of synergy; pathogen alone elicited strong responses including bZIP60 upregulation, consistent with greater aggressiveness on HR. The exogenous ABA experiment functionally supported the transcriptomic inference by reproducing enhanced infection and bZIP60 upregulation, reinforcing ABA/bZIP60 as key nodes in the synergy. While photosynthesis suppression via ACCase inhibition and possible thylakoid membrane destabilization may contribute to stress signaling, the primary mechanistic link to synergy appears to be ABA-activated signaling converging on bZIP60/UPR and protein ubiquitination pathways that compromise defense during P. setariae infection.

This work establishes a transcriptome-informed mechanism for the synergy between sublethal sethoxydim and the mycoherbicide Pyricularia setariae in herbicide-sensitive green foxtail. De novo transcriptomes for HS and HR biotypes enabled RNA-seq analyses showing that, specifically in HS, herbicide+pathogen co-treatment up-regulates ABA-activated signaling and bZIP60 (IRE1/UPR branch) and protein ubiquitination, while suppressing photosynthesis-related processes, leading to impaired defense and enhanced fungal biocontrol. Exogenous ABA phenocopied the synergistic effect and induced bZIP60, indicating a shared pathway. These findings provide a molecular basis for chemical–biological synergy in weed control and a resource for studying weed–pathogen interactions. Future work should: (1) functionally validate ABA/bZIP60 roles using genetic or chemical perturbations; (2) dissect ubiquitination components involved; (3) explore translation to field conditions and other weed–mycoherbicide systems; and (4) investigate mechanisms of herbicide resistance in HR GFT and strategies to overcome it.

- Sequencing depth and read length were limited by the MiSeq platform (~70M reads per biotype; moderate N50), potentially missing low-abundance transcripts and isoforms. - Only a single early infection time point (2 dpi) was profiled for transcriptomics; temporal dynamics of signaling and defense were not captured. - The resistance mechanism in the HR biotype was not determined, limiting interpretation of its transcriptional insensitivity to sethoxydim. - Experiments were conducted under controlled greenhouse conditions; field variability and environmental interactions were not assessed. - Functional roles of ABA/bZIP60 and ubiquitination were inferred from correlation and ABA application; direct causal validation (e.g., gene knockdown/overexpression) was not performed.